To get at the true state of American political parties, U. of Michigan political scientist Eldersveld marshals a mass of survey data--there are diagrams or charts on almost every page--and comes up with some unsurprising, generally wishy-washy conclusions. In sum: people who are involved are involved; people on the outside have a negative image from the media--so parties should communicate better. The evidence is presented in such forms as the ""Communication Model of Political Campaigning""--in which we see the ""Penetration,"" ""Boomerang,"" and ""Bypass"" effects graphically illustrated (penetration is good, the other two are bad). Discussing party recruitment of both cadres and candidates, Eldersveld winds up saying, typically, that ""although the evidence is not all in or crystal clear, there is evidence that the party organization can, and does, have a considerable role in the recruitment of leadership."" From all his data, Eldersveld decides that the parties are stronger at the local level--based on recruitment figures and strength of party identification (among other signs)--than most people, infected by the media, seem to think. In order to overcome the bad image, he proposes the following ""reform 'directions'"": more coherent party leadership; more deliberative and effective communicators; local organizations ""more societally relevant."" The image, alas, of an academic exercise.